1: \begin{definition}

2: \label{def_semi-Voronoi-reduced}

3: %%$\vee_{k=1}^{l}g_{m,k}(\tilde{y})$

4: Let $\B$ be a basis of $\Lambda$. Assume that  $\B$ and $\{ e_i \}_{i=1}^n$ have the same

5: orientation as in Theorem~\ref{th_func_VR}. The basis is called

6: semi-Voronoi-reduced (SVR) if there exists at least two points $x_1, x_2 \in \CBONE$

7: such that 

8: $\text{Surf}(\vee_{k=1}^{\ell_1}g_{1,k}) \bigcap \text{Surf}(\vee_{k=1}^{\ell_2}g_{2,k})

9: \ne \varnothing$, where $\ell_1,\ell_2\geq 1$,

10: $g_{1,k}$ are the facets between $x_1$ and all points in $\CN(x_1) \cap \CBZERO$,

11: and $g_{2,k}$ are the facets between $x_2$ and all points in $\CN(x_2) \cap \CBZERO$.

12: \end{definition}

13: